Though Florian Schwarz is a linguist, he pays a lot of attention to the twinkle in his research subjects’ eyes. That “twinkle” is the corneal reflection, or the spot where light reflects in the eye. The way it interacts with the pupil can teach scientists about the thought processes of the decision-makers they’re studying.

“The relative position of those two to one another changes as you move your eyes,” says Schwarz, an assistant professor in the Department of Linguistics in the School of Arts & Sciences and a member of the Psychology Graduate Group.

Schwarz studies meaning and natural language, comparing the difference between semantics, what a word means literally, and pragmatics or humans’ subconscious additions to that definition when using the word in context. The problem is, Schwarz says, when he asks people why they inferred a particular meaning, they tend to overthink it.

“Eye movements, however, are not subject to our explicit cognitive control,” he says, “so they give you insight into what the underlying cognitive processes actually are.”

Schwarz recently completed an eye-tracking study in his lab focusing on the word “some.” To understand the work requires a brief foray into Linguistics 101.

The word “some” can mean different things to different people. Most people would define it as meaning “some, but not all.” A linguist, Schwarz says, would counter that the term’s literal meaning is “some, and possibly all.” He says individuals add the “…but not all” part subconsciously.

“If you ask me, ‘Did you try some of Sue’s dishes?’ and I say, ‘Yes, I tried all of them,’ I’m not contradicting myself,” Schwarz says. “But if the question literally meant, ‘Did you try some but not all of those dishes?’ and I said, ‘Yes, I tried all of them,’ that doesn’t make any sense. Nonetheless, we usually understand ‘some’ as ‘some, but not all.’” Schwarz is trying to reconcile these two facts.

To test the notion, 78 Penn undergraduates viewed three images on a computer screen: a target (six cartoon giraffes with varying numbers wearing green scarves), a distractor (those six giraffes with no scarves), and a black box representing a hidden image. The students then heard the sentence, “Some of the giraffes have scarves.”

Participants clicked on the image best representing the statement, and the researchers tracked the subjects’ eye movement as the sentence unfolded.

“Even before hearing ‘scarves,’ people overwhelmingly shifted their attention to the picture where only some of the giraffes are wearing scarves,” Schwarz says. “They click on it without any second thoughts.” 

When the target picture had all six animals in scarves, however, volunteers’ responses were mixed. More selected the hidden picture represented by the black box despite not knowing its contents, and those who chose the target image took longer to do so.

“If they wound up accepting the target in the end, they kind of meandered to get there,” he says.

This tells Schwarz and other linguists that, in principle, people consider both the literal and pragmatic meanings of words regardless of their final response, but they more naturally use the pragmatic definition. It also suggests that the two meanings may compete with one another.

Research in this area is opening up a window into the closed-off brain.

“This kind of basic foundational research helps us understand the workings of the human mind,” Schwarz says.